Need Help creating models in PyMC3 as a beginner

Questions
1

Hello Everyone!

I am an undergrad student, and a complete beginner to probabilistic programming in general, so I am having some problems.

For the purpose of my dissertation project, I need to evaluate the efficacy of a certain vaccine against Covid-19 using Bayesian methods. As it turns out, my population parameter (infection process, i.e., number of people infected) follows a Poisson prior, and the mean of this prior, Lambda follows a gamma distribution. The likelihood turns out to be a binomial distribution.

At the present i am trying to model the posterior distribution of Lambda. The vaccine efficacy is a function of Lambda.

Can somebody guide me as to how I can create a model in PyMC3 for this?

As an added layer of complexity, I actually have two infection processes, one for the vaccine cohort and another for the control cohort. Can I get some hint as to how to proceed for this as well?

Thank you for patiently reading and answering!!

2

Hi,

you can go through this video and practice with the available notebook.
I found it quite useful.

All the best

3

Thanks!! I will definitely go through this!

Do you know of some other good, but lesser well known resources that i may use?

4

Hi there,
The examples on the pymc-dev repo are pretty neat. Here is the link for the repo implementing code examples from ā€˜Statistical Rethinkingā€™ textbook.

Some of the examples are explained with background - some not. Perhaps worth a (long-ish) shot at least?

1 Like
5

Yes will look through this definitely!!

thanks!